[Axiom-developer] Re: trunk

[Gabriel Dos Reis]

On Mon, 4 Jun 2007, address@hidden wrote:

| > I'm planning to attempt the following -- assuming I got time today:
| 
| 
| 
| >  * merge branches/daly to trunk
| 
| As I understand it trunk is Gold (aka --patch-50).

No, it wasn't.  I checked.  Many changes there were not present in
branches/daly. 

| But I wouldn't recommend doing that yet. There are more changes in
| the pipeline for branches/daly which are not fully merged and tested
| (e.g. bug 312 change, file-downcasing, fixes for .input files, Martin's
| guess algebra, etc). But you can do it if you like. I'd be just as
| happy to patch trunk as to patch branches/daly.

That is great.  What I would propose is that one you commit those changes to
branches/daly, we sync it with trunk and we continue with trunk.  That way,
you pending work is not disrupted and people are given sufficient notice of
what will happen.

| I think that it would be more fruitful if you spent time creating
| diff-Naur changesets between trunk and build-improvements or
| branches/daly and build-improvements.

I *did spend* time on that issue.  Just diffing between both branches is not
an option.  Changes I've made are buring deep into your and it just sucked a
fair amount of time (last week).  The fact that you did pick changes by
discrete diff/patch does not help.  I don't have much confience in manual
diffing. 

[...]

| >  * create a branch gold that contains the mature Gold version
| 
| Frankly, I'm not even sure what this would mean. But I am sure that
| creating a "branch" named "Gold" in SVN is certain to cause confusion
| and havoc. Gold lives in Arch, currently --patch-50. 

Having many versions of Axiom floating around causes greater confusion.
If people can get silver from one place, they should also be able to get gold
from that same place.

[...]

| I'm happy to commit to trunk rather than branches/daly.

That would be great.  Before that, please could you build trunk as of today
and run your regress script and tell me if something is wrong.
Ideally, we should be able to set the regress script so that it works for
everybody and the results sent to address@hidden

-- Gaby